Visualization In Python With Matplotlib

Visualization in Python with matplotlibPete AlonziResearch Data Services UVa LibraryMarch 1, 2016

Check us outdata.library.virginia.edu

What we’re gonna do today 2-3:30Get everyone up and running with pythonpython interfacesLoading matplotlibPlotting 101 (create and save)Histograms, scatter plotsCosmeticsmultiplots

Installation of Python Two Notes– Language (python) vs language distribution (anaconda)– Python 2 vs Python 3 https://www.continuum.io/downloads1. Scroll to your operating system2. Click on the blue box under Python 2.73. Follow the instructions

Interfaces Command line interpreter (and ipython)Commandline script executionIpython notebook / jupyterSpyder (generalize to ide)

Loading matplotlib It depends on your interface

A basic plot C:\Users\lpa2a python import matplotlib.pyplot as pltx range(10)plt.plot(x)plt.show()Why blue?Why a line?Why a line with slope of 1?Why does it hang after show?

Saving a plot There are two ways to save the plot:– Use the command line: plt.savefig(‘test.pdf’) Must be done before the show command– Use the gui: Click the save icon Type in the file name select the file type

Window control Notice when we use plt.show() we losecontrol. To regain control we must close the plotwindow. Now we’ll go through other ways ofinteracting with matplotlib to avoid thisproblem.

IPython %pylab Instead of loading up python at the command line withpython use ipython instead. Ipython has a special plotting mode which you load byissuing the command %pylab C:\Users\lpa2a ipython In [1]: %pylab Now we can try our basic plot again.––––Don’t need to load matplotliabWe don’t need to use the “plt.”We don’t loose control when we plotPlot appears on plot command, no more show()

Command line scrip execution Let’s take a look at C:\Users\lpa2a\plot.py

Ipython Notebook Launch from start menu Click “New Notebook”– In []: %pylab– Or– In []: %pylab inline

Spyder Runs the ipython interpreter– Use %pylab or import– Nb: cannot use “Run File” with %pylab

A quick note on appearance In the active plot window we observe a grayborder In the saved plot we observe a white border

Plot Generators There are a few functions in matplotlib thatwill cause a plot to be generated. So far we have worked with plot( ). Now we’ll look at a couple more– hist( )– scatter( )

Histograms

Histogram To plot a histogram we don’t use the function plot. Weuse the function hist––––import numpy as npimport matplotlib.pyplot as pltplt.hist(np.random.randn(1000))plt.show() All of the tricks we just learned to manipulate the plotstill work Here’s some examples for the binning– plt.hist(np.random.randn(1000),bins 25)– plt.hist(np.random.randn(1000),bins [5,-4,-3,-2,-1,0,1,2,3,4,5])17

Scatter Plot Use the function scatter()– 1000))– If using pylab– scatter(randn(1000),randn(1000))18

plot( ) So far we have used a simple implementationof plot. Let’s look deeper. %pylabplot(range(10)) #generatesclf()x arange(0,2*pi,0.2)plot(x,sin(x))x values

Multiplots To add multiple plots repeat the plot call– %pylab– x arange(0,2*pi,0.2)– plot1 plot(x,sin(x))– plot2 plot(x,cos(x)) Now to add a legend– plot1 plot(x,sin(x),label ‘sin’)– plot2 plot(x,cos(x),label ‘cos’)– legend(loc ‘best’)20

Subplots Multiple plots in the same figurePlotfigure21

Subplots For better control we will explicitly catch ourobjects––––––%pylabfig figure()sub1 fig.add subplot(2,2,1)sub2 fig.add subplot(2,2,3)plt.plot(arange(10))Use subplot(221) to switch active plot (demo) There is a function to do it all at once– fig,subs plt.subplots(3,3)Nb: need plt.show() on this one22

Let’s make our plot presentable C:\users\lpa2a ipython --pylab In[1]: plot(cos(arange(0,2*pi,0.2))) Grey backgroundAxis labels too smallPlot touches axisPlot not centered on axisHorizontal axis values aren’t what we wantNo axis labelsLine thicknessLine style23

Ranges and Values Set axis range– axis([-5,37,-1.5,1.5]) Change horizontal axis values– x arange(0,2*pi,0.2)– y cos(x)– plot(x,y)24

Labels and LaTeX Set axis labels– xlabel(‘x’,fontsize 20)– ylabel(‘cos(x)’)– title(‘Cosine’) You can use LaTeX as well– title(r’ \cos(x) ’) http://matplotlib.org/users/pyplot tutorial.html25

Linestyles You have a lot of freedom in choosing a linestyle. They can be expressed explicitly– plot(x,linestyle ‘--’) The same goes for line color– plot(x,color ‘g’) But you can also use shorthand– plot(x,’g--’)26

Linestyles II matplotlib automatically interpolates betweenthe points and puts in a line. To emphasize thepoints you can add markers.– plot(range(10),’o’) # markers, no line– plot(range(10),’o-’) # markers, line– plot(range(10),marker ‘o’) Set line thickness– pl plot(arange(10))– setp(pl,linewidth 5)27

One Page to Rule them All http://matplotlib.org/api/pyplot api.html– Comprehensive– Navigate with searching– Eg: ctrl f “.plot(“

Questions

Visualization in Python with matplotlib Pete Alonzi Research Data Services UVa Library March 1, 2016. Check us out data.library.virginia.edu. What we’re gonnado today 2-3:30 Get everyone up and running with python python interfaces Loading matplotlib Plotting 101 (create and save)